Public Lecture #2

"Semiconductor Technology – the Path Forward for the Coming Decades"

Prof. Philip WONG (Stanford)

Abstract

The semiconductor industry has been extremely successful in integrating discrete components into billion-transistor chips. Future electronic systems will continue to rely on, and increasingly benefit from, the advances in semiconductor technology as they have had for more than five decades. Applications such as AI, machine learning, 5G, and even quantum computing, will not fulfill their promises without the continual advancements of semiconductor technology that is anticipated.

21st century applications are going to be data-centric. Data analytics, machine learning, and AI applications are going to dominate, from data center to mobile and IoT, from collecting and processing, to curating the data to derive information. Many systems will need to learn and adapt on the fly.

The emergence of abundant-data computing made the system throughput and system throughput/Watt the key performance metrics. These metrics can be improved by more and more specialization from CPU to GPU, to TPU and accelerators that are able to execute a narrow set of tasks in a massively parallel fashion. In other words flexibility has been traded off to maximize system throughput and energy efficiency. Computation throughput has been advancing faster than the memory bandwidth, resulting in a bandwidth deficit that limits system performance.

With this in mind, this presentation will look at how the performance of those specialized architectures can be improved at the system level by advances of the underlying device and process technologies. An analysis of the technology trend for GPU and accelerators leads to a key observation: in the coming decades, we must go beyond a single chip from a wafer and focus on integrating chips into systems.

References:

1. M. M. S. Aly, … H.-S. P. Wong, S. Mitra, “The N3XT Approach to Energy-Efficient Abundant-Data Computing,” invited paper, Proceedings of the IEEE, vol. 107, pp. 19 – 48 (2018)
2. H. -S. P. Wong, et al. "A Density Metric for Semiconductor Technology," Proceedings of the IEEE, vol. 108, no. 4, pp. 478-482, April 2020
3. H.-S. P. Wong, “The Future is System Integration,” invited keynote, DARPA ERI Summit (ERI Summit), Seattle, WA, August 18 – 20, 2020. Youtube: https://youtu.be/cJh7EMWa6J0
4. H.-S. P. Wong, “Technology from 3D ICs to 3D System-on-a-Chip,” invited plenary paper, Design Automation Conference (DAC), San Francisco, July 20 – 22, 2020.  
5. H.-S. P. Wong, “IC Technology – What Will the Next Node Offer Us?” invited plenary paper, Hot Chips: A Symposium on High Performance Chips, Stanford, CA, August 18 – 20, 2019.  

Back